Evolving inborn knowledge for fast adaptation in dynamic POMDP problems
conference contributionposted on 11.05.2020 by Eseoghene Ben-Iwhiwhu, Pawel Ladosz, Jeff Dick, Wen-Hua Chen, Praveen Pilly, Andrea Soltoggio
Any type of content contributed to an academic conference, such as papers, presentations, lectures or proceedings.
Rapid online adaptation to changing tasks is an important problem in machine learning and, recently, a focus of meta-reinforcement learning. However, reinforcement learning (RL) algorithms struggle in POMDP environments because the state of the system, essential in a RL framework, is not always visible. Additionally, hand-designed meta-RL architectures may not include suitable computational structures for specific learning problems. The evolution of online learning mechanisms, on the contrary, has the ability to incorporate learning strategies into an agent that can (i) evolve memory when required and (ii) optimize adaptation speed to specific online learning problems. In this paper, we exploit the highly adaptive nature of neuromodulated neural networks to evolve a controller that uses the latent space of an autoencoder in a POMDP. The analysis of the evolved networks reveals the ability of the proposed algorithm to acquire inborn knowledge in a variety of aspects such as the detection of cues that reveal implicit rewards, and the ability to evolve location neurons that help with navigation. The integration of inborn knowledge and online plasticity enabled fast adaptation and better performance in comparison to some non-evolutionary meta-reinforcement learning algorithms. The algorithm proved also to succeed in the 3D gaming environment Malmo Minecraft.
Read the paper on the publisher website
United States Air Force Research Laboratory (AFRL) and Defense Advanced Research Projects Agency (DARPA) under Contract No. FA8750-18-C0103
- Aeronautical, Automotive, Chemical and Materials Engineering
- Aeronautical and Automotive Engineering
- Computer Science